{"id":"c376303d-a5b4-4b90-ae96-6f3c4c67d0eb","arxiv_id":"2501.15353","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A photonic chip measures rapid laser frequency changes in real time, resolving 2 MHz variations over a 160 nm range at speeds up to 2500 THz/s.","lead":"Researchers built a fingernail-sized photonic chip that measures fast changes in laser frequency in real time, down to about 2 MHz resolution and up to 2500 THz/s speed. If the results hold, the chip could shrink the frequency-monitoring hardware used in fiber sensors, LiDAR, and optical computing onto a single package.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed 2 MHz resolution and detected scan ripples rest on Eqs. S1–S4 assuming pure sinusoidal PD outputs; the paper's own Discussion concedes unmodeled back-reflection ripples, so the resolution could be a device artifact.","rationale":"The reader's weakest assumption is precisely where the central claim is least secure: the demodulation model assumes each detector output is a sinusoid whose only imperfections are bias, amplitude, modulation depth, and phase errors, all calibratable from Lissajous figures. The paper itself flags back-reflection-induced interferometric ripples as a degradation mechanism, and those ripples are not in the model. Because the 2 MHz resolution is inferred from filtered single traces, and the detected laser ripples are validated only by repeating the same laser scan, systematic device artifacts would also appear repeatable. The concern is not that the authors are wrong; the approach is plausible and the calibration is thoughtful. The concern is that the headline quantitative claims are not yet separated from possible unmodeled distortion. A direct noise-floor measurement with a stable laser, plus a perturbation test that changes the PD back-reflection environment, would settle the issue. If the residual noise stays below 2 MHz and the Fig. 4 ripples are unchanged after reflection suppression, the central claim is supported; if not, the resolution and ripple-detection claims would need to be revised. This is an addressable experimental gap, not an internal contradiction, so the appropriate verdict remains CONDITIONAL as the reader concluded.","tokens_in":21325,"tokens_out":3732,"duration_ms":39291,"concrete_test":"Feed an unmodulated narrow-linewidth reference laser (e.g., the NKT BASIK E15 with FM off, or a cavity-stabilized laser with sub-kHz linewidth) into the OFD and record the demodulated frequency for at least 10 s at the same 5 Hz post-filter bandwidth used for the 2 MHz claim; the residual standard deviation and peak-to-peak variation must be at or below 2 MHz. Then repeat the Fig. 4 scan after suppressing photodetector back reflections, for example by using index-matching adhesive at the butt-coupled interface or replacing the butt-coupled PDs with anti-reflection-angled or heterogeneously integrated PDs; if the 4.8 ms/0.93 ms ripple pattern changes or disappears, the claimed 'laser tuning ripple' and the 2 MHz resolution are contaminated by unmodeled device distortion.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The demodulation is built entirely on Eqs. S1–S4, which model each photodetector output as a single sinusoid in phase with only bias, amplitude, modulation-depth, and phase imperfections, all captured by Lissajous elliptical fits in SI Section 2. Any distortion not of this exact form is invisible to calibration: it is either absorbed into the fitted parameters or left as a residual phase error that masquerades as optical frequency variation. The paper's Discussion explicitly warns that back reflection from adhesively butt-coupled InGaAs photodetectors 'may cause back reflection induced interferometric ripples to degrade the OFD performance.' Such ripples are exactly the kind of non-sinusoidal, wavelength-dependent perturbation that Eqs. S1–S4 cannot represent. The repeatability argument in Fig. 4 does not rule this out: repeating the same laser scan three times repeats any OFD artifact tied to wavelength or scan position as well as any true laser ripple. Therefore the headline resolution of 2 MHz, the identification of 'tuning ripples not detectable with any other method,' and the derived FBG strain resolution all depend on an unverified assumption that the detector outputs are pure sinusoids. If back-reflection ripples are present, the demodulated phase, and hence every frequency trace in Figs. 3–7, is contaminated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a photonic integrated circuit (PIC) on thin-film lithium niobate (TFLN) implementing a sine-cosine optical frequency detector (OFD). The device consists of two unbalanced I-Q interferometers: a main interferometer with a 30 GHz FSR for high-resolution incremental frequency measurement and an assistive interferometer with a large FSR for absolute wavelength estimation. Calibration lookup tables capture wavelength-dependent circuit bias, amplitude, modulation depth, and phase imperfections. The authors demonstrate real-time characterization of a tunable laser's wavelength scans, detection of scan ripples and step transients, high-speed measurement up to 2500 THz/s, a claimed resolution down to 2 MHz (0.016 pm) at 5 Hz bandwidth, and an FBG interrogation application with strain resolution 0.1–0.2 µε at 500 Hz.","tokens_in":21553,"tokens_out":2904,"duration_ms":28061,"significance":"If the claims hold, the work is significant: it brings real-time, high-resolution optical frequency variation detection onto a compact chip, with broad wavelength coverage (1480–1640 nm) and potential for k-clock generation, laser frequency monitoring, and FBG interrogation. The paper's strengths include a clearly described demodulation formalism with explicit treatment of device imperfections, a standard Lissajous-fit calibration procedure, a wide operating range, and practical demonstrations with commercial lasers and an FBG strain sensor. The central limitation is that the headline resolution is inferred from visually discerned structure in filtered data rather than from a formal resolution metric, and the demodulation model assumes perfectly sinusoidal detector outputs that the paper's own discussion suggests may be violated.","major_comments":[{"comment":"The claim of a demonstrated resolution down to 2 MHz (0.016 pm) is not supported by a formal resolution metric. Fig. 6b shows a 10 MHz amplitude, 1 Hz sinusoidal modulation after digital low-pass filtering at 5 Hz; the 2 MHz figure is inferred from visual discernibility of the residual structure. No Allan deviation, noise-floor measurement, or statistical detection criterion is provided. Equations (1) and (2) are theoretical formulas with parameters that are not measured in this work. A resolution claim of this magnitude should be established with a proper measurement, such as Allan deviation of the demodulated phase with a stable input, or detection tests of known small frequency steps.","section":"High-resolution OFV measurements; Fig. 6b; Discussion"},{"comment":"The demodulation is built entirely on the assumption that each photodetector output is a single sinusoid in phase, with all imperfections captured by the calibration parameters (bias, amplitude, modulation depth, and phase errors). The paper's Discussion explicitly states that back reflection from adhesively butt-coupled InGaAs photodetectors 'may cause back reflection induced interferometric ripples to degrade the OFD performance.' Such ripples are non-sinusoidal and wavelength-dependent, and are not represented in Eqs. (S1)-(S4). If they are present, they will either be absorbed into the fitted calibration parameters or appear as residual phase errors that masquerade as optical frequency variations. The repeatability argument in Fig. 4 does not rule this out, because repeating the same laser scan repeats any artifact tied to wavelength or scan position. The authors should provide a direct test: for example, demodulate a laser whose frequency is independently known to be stable (e.g., via a beat measurement) and compare, or intentionally vary the PD reflection/input polarization to show the demodulated frequency is unchanged.","section":"Supplementary Eqs. (S1)-(S4), (S9); Discussion on back reflection"},{"comment":"The assistive (lower) interferometer is calibrated by assuming that its phase imperfection parameters α, β, γ are exactly the same as those of the main interferometer because the components are 'identical in design and fabrication,' and that B_lj has the same wavelength response up to a constant factor. This assumption is not verified experimentally. Because the assistive interferometer provides the absolute wavelength estimate used to select wavelength-dependent lookup-table parameters for the main interferometer, a mismatch could introduce systematic errors in the corrected demodulation. The authors should quantify the sensitivity of the final Δf(t) to plausible deviations in the assistive parameters, or perform a direct calibration of the lower interferometer without the identity assumption.","section":"Supplementary Section 3: Assistive interferometer calibration"}],"minor_comments":[{"comment":"The symbol τ_u is used for the time delay in Eq. (S5), but the text also uses π_u in several places; please make the notation consistent throughout.","section":"Eq. (S5) and surrounding text"},{"comment":"The symbol B is used for detection bandwidth in Eq. (3), while B_u1 etc. denote signal amplitudes in Eqs. (S1)-(S4); please disambiguate these uses.","section":"Eq. (3) and Table 1"},{"comment":"The caption states that the data were filtered with a second-order Butterworth low-pass filter at 200 kHz, but the figure shows only the filtered result. Showing an unfiltered segment would help the reader assess the raw noise level.","section":"Fig. 5d"},{"comment":"The strain resolution '0.1–0.2 µε at 500 Hz' is described as discerned from the figure; a quantitative criterion (e.g., standard deviation of the noise floor in the demodulated strain) should be reported.","section":"Fig. 7d and Discussion"},{"comment":"The paper contains several typographical issues (for example, 'FSRU' and 'SNRV' in Eqs. (1)-(4), and 'absolution' in SI Section 3); a careful proofreading pass is recommended.","section":"General presentation"}],"recommendation":"major_revision","confidential_remarks":"The work builds directly on the authors' prior sine-cosine encoder publications, but the PIC integration and the wide-wavelength calibration are new contributions. The paper's scope fits the journal, though the resolution claim and the model-assumption issue need to be addressed before publication. I would not recommend reject, as the concerns are addressable with additional measurements and analysis."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know: this paper reports the first TFLN photonic integrated circuit implementing Yao's sine-cosine optical frequency detector. That is a real step from the discrete-component demonstrations in the same group. The chip is compact (5.5 mm × 2.7 mm), operates over 1480–1640 nm, and the calibration procedure—Lissajous fits at 1 nm steps, a lookup table, and a low-FSR assistive interferometer for absolute wavelength—is clearly described and sensible. The demodulation math (Eqs. S1–S11) is internally consistent, and the authors are upfront about device imperfections and packaging limitations.\n\nThe main weakness is the resolution claim. The 2 MHz figure is asserted from visually discerning a 10 MHz modulation at 50 Hz bandwidth, after filtering; there is no Allan deviation or formal noise-floor measurement. Similarly, the 2500 THz/s speed is demonstrated by scanning a fast laser, but there is no independent reference to verify the measured chirp rate. The comparison to commercial FBG interrogators mixes measured strain resolution (0.1–0.2 µε at 500 Hz) with calculated values derived from the assumed 2 MHz resolution.\n\nThe stress-test concern is legitimate: the demodulation model assumes each photodetector output is a single sinusoid with only bias, amplitude, phase, and modulation-depth errors. The Discussion explicitly concedes that back-reflection from the butt-coupled PDs can cause interferometric ripples, which would violate that model and appear as false frequency ripples. The repeatability argument does not rule this out, since a wavelength-dependent artifact would be repeatable. That is not a fatal flaw—the calibration is standard and the device concept is sound—but it means the headline resolution and the interpretation of 'tuning ripples' are not fully established.\n\nCredit where due: the paper does not hide these issues. It lists them as limitations and suggests concrete fixes (heterogeneous PD integration, edge coupling, better TIA placement). The core contribution—an integrated, calibrated sine-cosine OFD—is genuinely useful and likely to be adopted by groups working on laser control, FMCW LiDAR, or FBG interrogation.\n\nI would engage with this. Send it to peer review, but ask the authors to add: (1) a formal resolution measurement (Allan deviation or noise-equivalent frequency), (2) an independent check of a high-speed scan against a second interferometer or wavemeter, and (3) a test of the back-reflection effect, e.g., by index-matching the PD coupling or repeating a measurement with a different coupling method. All three are addressable in a revision.","headline":"A solid integrated-photonics demonstration of a sine-cosine optical frequency detector on TFLN, with a sensible calibration scheme and honest limitations, but the headline 2 MHz resolution is inferred from filtered traces rather than a formal noise characterization, and back-reflection artifacts could contaminate the demodulation.","tokens_in":22185,"tokens_out":2369,"would_cite":true,"duration_ms":21879,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A 5.5-mm photonic chip detects laser frequency changes as small as 2 MHz in real time while tracking sweeps up to 2500 THz/s.","keywords":["optical frequency variation","sine-cosine encoder","thin-film lithium niobate","photonic integrated circuit","I-Q interferometer","FBG interrogation","real-time laser characterization","k-clock generation"],"falsifier":"Measure a laser whose frequency is independently known to be stable to below 1 MHz over several seconds, using a dual-comb or cavity-stabilized reference, while feeding the chip the same light. If the chip reports periodic ripples at its own free spectral range, at the detector cavity round-trip frequency, or at the 50-Hz power-line harmonics that are absent in the reference, the single-sinusoid calibration model is incomplete and the claimed 2 MHz resolution would be an artifact of the model rather than a true optical frequency measurement.","tokens_in":21087,"feed_emoji":"📡","tokens_out":4410,"duration_ms":41825,"temperature":0.7,"pith_summary":"This paper reports a chip-sized optical frequency detector that turns the sine-cosine encoder principle, familiar from motor position sensing, into an on-chip interferometric measurement of laser frequency variation. The central claim is that a 5.5 mm by 2.7 mm thin-film lithium niobate chip can detect frequency changes down to 2 MHz (0.016 pm) and follow frequency sweeps as fast as 2500 THz/s across a 160-nm wavelength range, with a demodulation algorithm that absorbs device imperfections through calibration. If true, this would replace bulky fiber interferometers, optical spectrum analyzers, and comb-based systems in applications needing real-time frequency tracking. The paper also demonstrates the chip as a fiber Bragg grating interrogator, resolving dynamic strain at 0.1 to 0.2 microstrain at 500 Hz, and argues that better electronics could push sensitivity well beyond current commercial interrogators.","feed_headline":"Photonic chip catches 2-MHz laser shifts in real time","feed_subtitle":"Two on-chip interferometers read laser frequency changes over a 160-nm range, sharpening strain and vibration sensing.","key_machinery":"The central object is a photonic sine-cosine encoder built from two unbalanced I-Q interferometers, each made of a coupler and a 90-degree hybrid realized as a 2x4 multimode interference coupler. The upper interferometer provides fine frequency increments through a small free spectral range, while the lower interferometer provides the absolute wavelength reference through a very large free spectral range. The demodulation formula, Eq. (S9), computes the frequency increment from the four normalized photodetector voltages using calibrated bias, amplitude, modulation-depth, and phase-imperfection parameters, so that wavelength-dependent fabrication deviations are corrected rather than treated as ideal.","core_discovery":"The authors demonstrate that two unbalanced I-Q interferometers on a single TFLN chip, one with a 10-mm optical path difference (30 GHz free spectral range) for fine frequency increments and one with a 14.43-µm optical path difference (20.79 THz free spectral range) for absolute wavelength estimation, can measure optical frequency variations in real time. The key to the claimed performance is a calibration and demodulation procedure that models each photodetector output as a sinusoid with independent bias, amplitude, modulation depth, and quadrature phase imperfections, all wavelength dependent, and then uses an arctangent formula to recover the phase increment and hence the frequency variation. With this approach, the authors show that they can resolve 1-pm wavelength steps, reveal scanning ripples and overshoots in commercial tunable lasers, measure a 100-kHz sinusoidal frequency modulation, and track a 20000-nm/s sweep with stair-step detail. They further show that the same chip can interrogate a fiber Bragg grating, extracting damped oscillations and 500-Hz acoustic strain at resolutions they state are roughly an order of magnitude finer than the best commercial interrogators.","pith_inferences":["The calibration model is general enough that the same demodulation should transfer to silicon or silicon-nitride platforms, provided the dispersion of the interferometer's free spectral range is measured; the paper's lookup-table approach does not depend on the TFLN material itself.","If butt-coupled InGaAs photodetectors were replaced by heterogeneously integrated photodetectors, as the paper suggests, the back-reflection-induced ripples it identifies as a performance limit could be eliminated, potentially bringing the resolution closer to the 458-kHz DAQ-limited value at 1 MHz bandwidth.","The demonstrated ability to resolve 1-pm laser steps and scan ripples suggests the chip could serve as a real-time diagnostic for laser tuning mechanisms, not just as a sensor; that diagnostic role is implied but not developed in the paper.","Adding a thermally tunable microring filter or arrayed waveguide grating in front of the OFD could enable real-time frequency monitoring of individual optical comb lines, an extension the paper sketches but does not demonstrate experimentally."],"forward_implications":["On-chip k-clock generation for OFDR, FMCW LiDAR, and OCT becomes feasible: the paper shows that an OPD of 300 m needed with a conventional unbalanced Mach-Zehnder interferometer could shrink to about 9.16 mm with the sine-cosine OFD, small enough to integrate on a sensing chip.","Chip-scale FBG interrogation becomes faster and more sensitive, with projected strain resolution of 0.013 microstrain and temperature resolution of 0.0015 degrees Celsius at 5 Hz bandwidth, improving further if lower-noise electronics are used.","Tunable laser sources can be characterized in real time with enough detail to reveal scan nonlinearity, ripple, step overshoot, and scan-rate irregularity, information that could feed closed-loop laser frequency control.","The same architecture could be extended to other wavelength bands because TFLN is transparent from 350 nm to 5500 nm, and longer optical path differences on low-loss SiN waveguides could push frequency resolution from 2 MHz toward tens of kilohertz.","The device can measure not just frequency but frequency variation rate, enabling detection of chirp rates up to 2500 THz/s with a simple 62.5-MS/s digitizer, which is relevant for FMCW laser chirp characterization."],"supporting_citations":[{"why":"Supplies the sine-cosine OFD principle and the resolution and speed equations (Eqs. 1-4) that set the 2 MHz and 2500 THz/s performance limits.","marker":"[19]"},{"why":"The C-band TFLN 90-degree hybrid is the quadrature component that produces the simultaneous sine and cosine signals.","marker":"[40]"},{"why":"Hilbert-transform processing of an unbalanced interferometer is described as non-real-time, motivating the need for the direct sine-cosine approach.","marker":"[21]"},{"why":"A fully integrated on-chip FBG interrogator serves as a baseline for comparison of wavelength resolution and interrogation mechanism.","marker":"[4]"},{"why":"Dual-comb characterization of 1500 THz/s chirps is the earlier high-speed benchmark against which the chip's speed performance is positioned.","marker":"[31]"},{"why":"Heterogeneously integrated InP/InGaAs photodetectors on TFLN are cited as the route to remove back-reflection-induced ripples that can degrade OFD performance.","marker":"[45]"},{"why":"A survey of photonic-integrated FBG interrogators provides the context for the claimed sensitivity and speed advantages.","marker":"[41]"}],"fun_headline_variants":["Chip-scale detector tracks laser frequency shifts in real time","Ultrafast photonic chip resolves 2-MHz optical changes on the fly","On-chip sine-cosine encoder measures frequency jitter at 2-MHz resolution","Tiny TFLN chip detects optical frequency changes at 2,500 THz/s","Photonic chip beats fiber sensors in speed and sensitivity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The demodulation model assumes each photodetector output is a single sinusoid whose deviations from ideal are fully captured by a handful of calibrated bias, amplitude, modulation-depth, and phase parameters, and the paper itself warns in the Discussion that back reflection from adhesively bonded photodetectors can add interferometric ripples that this model does not represent.","fun_headline_variants_meta":{"raw":{"variants":["Chip-scale detector tracks laser frequency shifts in real time","Ultrafast photonic chip resolves 2-MHz optical changes on the fly","On-chip sine-cosine encoder measures frequency jitter at 2-MHz resolution","Tiny TFLN chip detects optical frequency changes at 2,500 THz/s","Photonic chip beats fiber sensors in speed and sensitivity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000311,"raw_usage":{"total_tokens":1786,"prompt_tokens":976,"completion_tokens":810,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":592,"completion_tokens_details":{"reasoning_tokens":713}},"tokens_in":592,"tokens_out":810,"duration_ms":7460,"temperature":1.0,"reasoning_tokens":713,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T14:22:17.730981+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure a laser whose frequency is independently known to be stable to below 1 MHz over several seconds, using a dual-comb or cavity-stabilized reference, while feeding the chip the same light. If the chip reports periodic ripples at its own free spectral range, at the detector cavity round-trip frequency, or at the 50-Hz power-line harmonics that are absent in the reference, the single-sinusoid calibration model is incomplete and the claimed 2 MHz resolution would be an artifact of the model rather than a true optical frequency measurement.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the sine-cosine OFD principle and the resolution and speed equations (Eqs. 1-4) that set the 2 MHz and 2500 THz/s performance limits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Hilbert-transform processing of an unbalanced interferometer is described as non-real-time, motivating the need for the direct sine-cosine approach."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"A fully integrated on-chip FBG interrogator serves as a baseline for comparison of wavelength resolution and interrogation mechanism."},{"cited_title":"R., Baumann, E., Swann, W","cited_arxiv_id":null,"evidence_quote":"Dual-comb characterization of 1500 THz/s chirps is the earlier high-speed benchmark against which the chip's speed performance is positioned."},{"cited_title":"Ultra-wideband waveguide-coupled photodiodes heterogeneously integrated on a thin-film lithium niobate platform","cited_arxiv_id":null,"evidence_quote":"Heterogeneously integrated InP/InGaAs photodetectors on TFLN are cited as the route to remove back-reflection-induced ripples that can degrade OFD performance."},{"cited_title":"E., Nannipieri, T., Oton, C","cited_arxiv_id":null,"evidence_quote":"A survey of photonic-integrated FBG interrogators provides the context for the claimed sensitivity and speed advantages."}],"review_version":1}